Autologous bone marrow transplantation is a potentially curative approach to refractory cancer, particularly in children. Its major limitation is the likelihood of occult bone marrow infiltration. We will use neuroblastoma as a model for developing ways of rendering bone marrow free of neoplastic cells with a high degree of certainty. Our approach is to use selective antitumor antibodies to identify neoplastic cells in bone marrow and then eliminate those cells by addition of complement. This represents the most physiological approach to the problem and is least likely to introduce undesirable products into the patient. The major problem in using antibody and complement for rendering bone marrow free of tumor cells is the presence of subpopulations of tumor cells resistant to complement action. The studies will focus on methods of manipulating the sensitivity of the tumor cells to complement killing to enable 100% of antibody-sensitized tumor cells to be destroyed. Methods that result in effective tumor cell elimination will be tested for effects on normal bone marrow stem cells, and procedures that result in tumor cell killing with minimal damage to normal stem cells will be identified. Finally, the practicality of scaling up these procedures to provide sufficient tumor-free bone marrow for autologous bone marrow transplantation will be tested to establish whether this approach is clinically applicable. The potential for ablative cytotoxic therapy supported by autologous bone marrow in the treatment of neuroblastoma cannot be realistically assessed until we can demonstrate reproducibly the elimination of neoplastic cells from bone marrow in vitro without damage to bone marrow stem cell populations. (IT)